Abstract:
Twitter is a social networking site that allows a large number of users to communicate with each
other. Twitter allows users to share their views on different topics ranging from day to day life to
what is going in society. As users are sharing what is going around them, this makes twitter a
good source of information. Twitter provides very less lag between the time events have
happened and when it is first reported on Twitter. Event detection in twitter is the process of
detecting popular events using messages generated by the users. Event detection is difficult in
twitter as compared to other media because the message known as tweets is only allowed to be
less than 140 characters. This means that the content in tweets will be much focused and may
contain short forms. Secondly the tweets are noisy because there may be personal messages by
the user also. We have designed an algorithm that finds top k popular events from tweets using
keywords contained in the tweets. This thesis also classified the popular events into different
categories. Analysis of events is also done in this thesis such as timeline interface and map. Event
analysis provides more insight into one particular event and helps to understand an event
thoroughly. The timeline is useful to check when the event was popular and map is useful to
check where the event was popular. To do all this experimental work we have chosen 14,558
users. We have collected 5,27,548 tweets over a period of 10 months (22 June, 2015 to 25 April,
2016).
